/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Diff of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1359 by ph10, Tue Sep 3 10:10:59 2013 UTC revision 1409 by ph10, Fri Dec 6 12:07:47 2013 UTC
# Line 1174  time, run time, or study time, respectiv Line 1174  time, run time, or study time, respectiv
1174  #define PUBLIC_COMPILE_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
1175    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1176     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1177     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \
1178       PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1179     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1180     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1181    
# Line 1532  a positive value. */ Line 1533  a positive value. */
1533  #define STRING_xdigit               "xdigit"  #define STRING_xdigit               "xdigit"
1534    
1535  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1536    #define STRING_WEIRD_STARTWORD      "[:<:]]"
1537    #define STRING_WEIRD_ENDWORD        "[:>:]]"
1538    
1539  #define STRING_CR_RIGHTPAR             "CR)"  #define STRING_CR_RIGHTPAR              "CR)"
1540  #define STRING_LF_RIGHTPAR             "LF)"  #define STRING_LF_RIGHTPAR              "LF)"
1541  #define STRING_CRLF_RIGHTPAR           "CRLF)"  #define STRING_CRLF_RIGHTPAR            "CRLF)"
1542  #define STRING_ANY_RIGHTPAR            "ANY)"  #define STRING_ANY_RIGHTPAR             "ANY)"
1543  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR         "ANYCRLF)"
1544  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR     "BSR_ANYCRLF)"
1545  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR     "BSR_UNICODE)"
1546  #define STRING_UTF8_RIGHTPAR           "UTF8)"  #define STRING_UTF8_RIGHTPAR            "UTF8)"
1547  #define STRING_UTF16_RIGHTPAR          "UTF16)"  #define STRING_UTF16_RIGHTPAR           "UTF16)"
1548  #define STRING_UTF32_RIGHTPAR          "UTF32)"  #define STRING_UTF32_RIGHTPAR           "UTF32)"
1549  #define STRING_UTF_RIGHTPAR            "UTF)"  #define STRING_UTF_RIGHTPAR             "UTF)"
1550  #define STRING_UCP_RIGHTPAR            "UCP)"  #define STRING_UCP_RIGHTPAR             "UCP)"
1551  #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"  #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1552  #define STRING_LIMIT_MATCH_EQ          "LIMIT_MATCH="  #define STRING_NO_START_OPT_RIGHTPAR    "NO_START_OPT)"
1553  #define STRING_LIMIT_RECURSION_EQ      "LIMIT_RECURSION="  #define STRING_LIMIT_MATCH_EQ           "LIMIT_MATCH="
1554    #define STRING_LIMIT_RECURSION_EQ       "LIMIT_RECURSION="
1555    
1556  #else  /* SUPPORT_UTF */  #else  /* SUPPORT_UTF */
1557    
# Line 1795  only. */ Line 1799  only. */
1799  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1800    
1801  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1802    #define STRING_WEIRD_STARTWORD      STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1803    #define STRING_WEIRD_ENDWORD        STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1804    
1805  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR              STR_C STR_R STR_RIGHT_PARENTHESIS
1806  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR              STR_L STR_F STR_RIGHT_PARENTHESIS
1807  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR            STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1808  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR             STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1809  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR         STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1810  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1811  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1812  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1813  #define STRING_UTF16_RIGHTPAR          STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS  #define STRING_UTF16_RIGHTPAR           STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1814  #define STRING_UTF32_RIGHTPAR          STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS  #define STRING_UTF32_RIGHTPAR           STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1815  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_RIGHT_PARENTHESIS  #define STRING_UTF_RIGHTPAR             STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1816  #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS  #define STRING_UCP_RIGHTPAR             STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1817  #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS  #define STRING_NO_AUTO_POSSESS_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_A STR_U STR_T STR_O STR_UNDERSCORE STR_P STR_O STR_S STR_S STR_E STR_S STR_S STR_RIGHT_PARENTHESIS
1818  #define STRING_LIMIT_MATCH_EQ          STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN  #define STRING_NO_START_OPT_RIGHTPAR    STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
1819  #define STRING_LIMIT_RECURSION_EQ      STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN  #define STRING_LIMIT_MATCH_EQ           STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN
1820    #define STRING_LIMIT_RECURSION_EQ       STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN
1821    
1822  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
1823    
# Line 1852  only. */ Line 1859  only. */
1859  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1860  #define PT_CLIST      9    /* Pseudo-property: match character list */  #define PT_CLIST      9    /* Pseudo-property: match character list */
1861  #define PT_UCNC      10    /* Universal Character nameable character */  #define PT_UCNC      10    /* Universal Character nameable character */
1862    #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
1863    
1864    /* The following special properties are used only in XCLASS items, when POSIX
1865    classes are specified and PCRE_UCP is set - in other words, for Unicode
1866    handling of these classes. They are not available via the \p or \P escapes like
1867    those in the above list, and so they do not take part in the autopossessifying
1868    table. */
1869    
1870    #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
1871    #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
1872    #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
1873    
1874  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1875  contain characters with values greater than 255. */  contain characters with values greater than 255. */
# Line 1866  contain characters with values greater t Line 1884  contain characters with values greater t
1884  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1885    
1886  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
1887  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns 0
1888  0 for a data character.  Also, they must appear in the same order as in the opcode  for a data character.  Also, they must appear in the same order as in the
1889  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it  opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1890  corresponds to "." in DOTALL mode rather than an escape sequence. It is also  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1891  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1892  non-DOTALL mode, "." behaves like \N.  non-DOTALL mode, "." behaves like \N.
# Line 1891  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1909  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1909         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1910         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1911    
 /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  
 OP_EOD must correspond in order to the list of escapes immediately above.  
1912    
1913  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  /********************** Opcode definitions ******************/
1914  that follow must also be updated to match. There are also tables called  
1915  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  /****** NOTE NOTE NOTE ******
1916    
1917    Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1918    order to the list of escapes immediately above. Furthermore, values up to
1919    OP_DOLLM must not be changed without adjusting the table called autoposstab in
1920    pcre_compile.c
1921    
1922    Whenever this list is updated, the two macro definitions that follow must be
1923    updated to match. The possessification table called "opcode_possessify" in
1924    pcre_compile.c must also be updated, and also the tables called "coptable"
1925    and "poptable" in pcre_dfa_exec.c.
1926    
1927    ****** NOTE NOTE NOTE ******/
1928    
1929    
1930    /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1931    are used in a table for deciding whether a repeated character type can be
1932    auto-possessified. */
1933    
1934    #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
1935    #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
1936    #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
1937    
1938  enum {  enum {
1939    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1929  enum { Line 1966  enum {
1966    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */
1967    OP_EOD,            /* 24 End of data (\z) */    OP_EOD,            /* 24 End of data (\z) */
1968    
1969    OP_CIRC,           /* 25 Start of line - not multiline */    /* Line end assertions */
1970    OP_CIRCM,          /* 26 Start of line - multiline */  
1971    OP_DOLL,           /* 27 End of line - not multiline */    OP_DOLL,           /* 25 End of line - not multiline */
1972    OP_DOLLM,          /* 28 End of line - multiline */    OP_DOLLM,          /* 26 End of line - multiline */
1973      OP_CIRC,           /* 27 Start of line - not multiline */
1974      OP_CIRCM,          /* 28 Start of line - multiline */
1975    
1976      /* Single characters; caseful must precede the caseless ones */
1977    
1978    OP_CHAR,           /* 29 Match one character, casefully */    OP_CHAR,           /* 29 Match one character, casefully */
1979    OP_CHARI,          /* 30 Match one character, caselessly */    OP_CHARI,          /* 30 Match one character, caselessly */
1980    OP_NOT,            /* 31 Match one character, not the given one, casefully */    OP_NOT,            /* 31 Match one character, not the given one, casefully */
# Line 1941  enum { Line 1983  enum {
1983    /* The following sets of 13 opcodes must always be kept in step because    /* The following sets of 13 opcodes must always be kept in step because
1984    the offset from the first one is used to generate the others. */    the offset from the first one is used to generate the others. */
1985    
1986    /**** Single characters, caseful, must precede the caseless ones ****/    /* Repeated characters; caseful must precede the caseless ones */
1987    
1988    OP_STAR,           /* 33 The maximizing and minimizing versions of */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1989    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
# Line 1959  enum { Line 2001  enum {
2001    OP_POSQUERY,       /* 44 Posesssified query, caseful */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
2002    OP_POSUPTO,        /* 45 Possessified upto, caseful */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
2003    
2004    /**** Single characters, caseless, must follow the caseful ones */    /* Repeated characters; caseless must follow the caseful ones */
2005    
2006    OP_STARI,          /* 46 */    OP_STARI,          /* 46 */
2007    OP_MINSTARI,       /* 47 */    OP_MINSTARI,       /* 47 */
# Line 1977  enum { Line 2019  enum {
2019    OP_POSQUERYI,      /* 57 Posesssified query, caseless */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
2020    OP_POSUPTOI,       /* 58 Possessified upto, caseless */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
2021    
2022    /**** The negated ones must follow the non-negated ones, and match them ****/    /* The negated ones must follow the non-negated ones, and match them */
2023    /**** Negated single character, caseful; must precede the caseless ones ****/    /* Negated repeated character, caseful; must precede the caseless ones */
2024    
2025    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
2026    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
# Line 1996  enum { Line 2038  enum {
2038    OP_NOTPOSQUERY,    /* 70 */    OP_NOTPOSQUERY,    /* 70 */
2039    OP_NOTPOSUPTO,     /* 71 */    OP_NOTPOSUPTO,     /* 71 */
2040    
2041    /**** Negated single character, caseless; must follow the caseful ones ****/    /* Negated repeated character, caseless; must follow the caseful ones */
2042    
2043    OP_NOTSTARI,       /* 72 */    OP_NOTSTARI,       /* 72 */
2044    OP_NOTMINSTARI,    /* 73 */    OP_NOTMINSTARI,    /* 73 */
# Line 2014  enum { Line 2056  enum {
2056    OP_NOTPOSQUERYI,   /* 83 */    OP_NOTPOSQUERYI,   /* 83 */
2057    OP_NOTPOSUPTOI,    /* 84 */    OP_NOTPOSUPTOI,    /* 84 */
2058    
2059    /**** Character types ****/    /* Character types */
2060    
2061    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
2062    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
# Line 2045  enum { Line 2087  enum {
2087    OP_CRRANGE,        /* 104 These are different to the three sets above. */    OP_CRRANGE,        /* 104 These are different to the three sets above. */
2088    OP_CRMINRANGE,     /* 105 */    OP_CRMINRANGE,     /* 105 */
2089    
2090      OP_CRPOSSTAR,      /* 106 Possessified versions */
2091      OP_CRPOSPLUS,      /* 107 */
2092      OP_CRPOSQUERY,     /* 108 */
2093      OP_CRPOSRANGE,     /* 109 */
2094    
2095    /* End of quantifier opcodes */    /* End of quantifier opcodes */
2096    
2097    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 110 Match a character class, chars < 256 only */
2098    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
2099                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2100                                character > 255 is encountered. */                                character > 255 is encountered. */
2101    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the    OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
2102                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2103    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 113 Match a back reference, casefully */
2104    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 114 Match a back reference, caselessly */
2105    OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */    OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
2106    OP_CALLOUT,        /* 112 Call out to external function if provided */    OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
2107      OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
2108    OP_ALT,            /* 113 Start of alternation */    OP_CALLOUT,        /* 118 Call out to external function if provided */
2109    OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */  
2110    OP_KETRMAX,        /* 115 These two must remain together and in this */    OP_ALT,            /* 119 Start of alternation */
2111    OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */    OP_KET,            /* 120 End of group that doesn't have an unbounded repeat */
2112    OP_KETRPOS,        /* 117 Possessive unlimited repeat. */    OP_KETRMAX,        /* 121 These two must remain together and in this */
2113      OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
2114      OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
2115    
2116    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2117    asserts must remain in order. */    asserts must remain in order. */
2118    
2119    OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
2120    OP_ASSERT,         /* 119 Positive lookahead */    OP_ASSERT,         /* 125 Positive lookahead */
2121    OP_ASSERT_NOT,     /* 120 Negative lookahead */    OP_ASSERT_NOT,     /* 126 Negative lookahead */
2122    OP_ASSERTBACK,     /* 121 Positive lookbehind */    OP_ASSERTBACK,     /* 127 Positive lookbehind */
2123    OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2124    
2125    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2126    after the assertions, with ONCE first, as there's a test for >= ONCE for a    after the assertions, with ONCE first, as there's a test for >= ONCE for a
2127    subpattern that isn't an assertion. The POS versions must immediately follow    subpattern that isn't an assertion. The POS versions must immediately follow
2128    the non-POS versions in each case. */    the non-POS versions in each case. */
2129    
2130    OP_ONCE,           /* 123 Atomic group, contains captures */    OP_ONCE,           /* 129 Atomic group, contains captures */
2131    OP_ONCE_NC,        /* 124 Atomic group containing no captures */    OP_ONCE_NC,        /* 130 Atomic group containing no captures */
2132    OP_BRA,            /* 125 Start of non-capturing bracket */    OP_BRA,            /* 131 Start of non-capturing bracket */
2133    OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */    OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
2134    OP_CBRA,           /* 127 Start of capturing bracket */    OP_CBRA,           /* 133 Start of capturing bracket */
2135    OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */    OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
2136    OP_COND,           /* 129 Conditional group */    OP_COND,           /* 135 Conditional group */
2137    
2138    /* These five must follow the previous five, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
2139    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
2140    
2141    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
2142    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */    OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
2143    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */    OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
2144    OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */    OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
2145    OP_SCOND,          /* 134 Conditional group, check empty */    OP_SCOND,          /* 140 Conditional group, check empty */
2146    
2147    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
2148    
2149    OP_CREF,           /* 135 Used to hold a capture number as condition */    OP_CREF,           /* 141 Used to hold a capture number as condition */
2150    OP_NCREF,          /* 136 Same, but generated by a name reference*/    OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
2151    OP_RREF,           /* 137 Used to hold a recursion number as condition */    OP_RREF,           /* 143 Used to hold a recursion number as condition */
2152    OP_NRREF,          /* 138 Same, but generated by a name reference*/    OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
2153    OP_DEF,            /* 139 The DEFINE condition */    OP_DEF,            /* 145 The DEFINE condition */
2154    
2155    OP_BRAZERO,        /* 140 These two must remain together and in this */    OP_BRAZERO,        /* 146 These two must remain together and in this */
2156    OP_BRAMINZERO,     /* 141 order. */    OP_BRAMINZERO,     /* 147 order. */
2157    OP_BRAPOSZERO,     /* 142 */    OP_BRAPOSZERO,     /* 148 */
2158    
2159    /* These are backtracking control verbs */    /* These are backtracking control verbs */
2160    
2161    OP_MARK,           /* 143 always has an argument */    OP_MARK,           /* 149 always has an argument */
2162    OP_PRUNE,          /* 144 */    OP_PRUNE,          /* 150 */
2163    OP_PRUNE_ARG,      /* 145 same, but with argument */    OP_PRUNE_ARG,      /* 151 same, but with argument */
2164    OP_SKIP,           /* 146 */    OP_SKIP,           /* 152 */
2165    OP_SKIP_ARG,       /* 147 same, but with argument */    OP_SKIP_ARG,       /* 153 same, but with argument */
2166    OP_THEN,           /* 148 */    OP_THEN,           /* 154 */
2167    OP_THEN_ARG,       /* 149 same, but with argument */    OP_THEN_ARG,       /* 155 same, but with argument */
2168    OP_COMMIT,         /* 150 */    OP_COMMIT,         /* 156 */
2169    
2170    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
2171    
2172    OP_FAIL,           /* 151 */    OP_FAIL,           /* 157 */
2173    OP_ACCEPT,         /* 152 */    OP_ACCEPT,         /* 158 */
2174    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */    OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
2175    OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
2176    
2177    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
2178    
2179    OP_SKIPZERO,       /* 155 */    OP_SKIPZERO,       /* 161 */
2180    
2181    /* This is not an opcode, but is used to check that tables indexed by opcode    /* This is not an opcode, but is used to check that tables indexed by opcode
2182    are the correct length, in order to catch updating errors - there have been    are the correct length, in order to catch updating errors - there have been
# Line 2138  enum { Line 2187  enum {
2187    
2188  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2189  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
2190  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
2191    pcre_dfa_exec.c that must be updated. */
2192    
2193    
2194  /* This macro defines textual names for all the opcodes. These are used only  /* This macro defines textual names for all the opcodes. These are used only
# Line 2151  some cases doesn't actually use these na Line 2201  some cases doesn't actually use these na
2201    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
2202    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
2203    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
2204    "^", "^", "$", "$", "char", "chari", "not", "noti",             \    "$", "$", "^", "^", "char", "chari", "not", "noti",             \
2205    "*", "*?", "+", "+?", "?", "??",                                \    "*", "*?", "+", "+?", "?", "??",                                \
2206    "{", "{", "{",                                                  \    "{", "{", "{",                                                  \
2207    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 2167  some cases doesn't actually use these na Line 2217  some cases doesn't actually use these na
2217    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
2218    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
2219    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
2220    "class", "nclass", "xclass", "Ref", "Refi",                     \    "*+","++", "?+", "{",                                           \
2221      "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
2222    "Recurse", "Callout",                                           \    "Recurse", "Callout",                                           \
2223    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
2224    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
# Line 2176  some cases doesn't actually use these na Line 2227  some cases doesn't actually use these na
2227    "Cond",                                                         \    "Cond",                                                         \
2228    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
2229    "SCond",                                                        \    "SCond",                                                        \
2230    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2231    "Brazero", "Braminzero", "Braposzero",                          \    "Brazero", "Braminzero", "Braposzero",                          \
2232    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
2233    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
# Line 2201  in UTF-8 mode. The code that uses this t Line 2252  in UTF-8 mode. The code that uses this t
2252    3, 3,                          /* \P, \p                                 */ \    3, 3,                          /* \P, \p                                 */ \
2253    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
2254    1,                             /* \X                                     */ \    1,                             /* \X                                     */ \
2255    1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \    1, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
2256    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
2257    2,                             /* Chari  - the minimum length            */ \    2,                             /* Chari  - the minimum length            */ \
2258    2,                             /* not                                    */ \    2,                             /* not                                    */ \
# Line 2232  in UTF-8 mode. The code that uses this t Line 2283  in UTF-8 mode. The code that uses this t
2283    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2284    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2285    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2286      1, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
2287    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2288    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2289    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2290    1+IMM2_SIZE,                   /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2291    1+IMM2_SIZE,                   /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2292      1+2*IMM2_SIZE,                 /* DNREF                                  */ \
2293      1+2*IMM2_SIZE,                 /* DNREFI                                 */ \
2294    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2295    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2296    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 2261  in UTF-8 mode. The code that uses this t Line 2315  in UTF-8 mode. The code that uses this t
2315    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2316    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2317    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2318    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
2319    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
2320    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2321    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2322    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
# Line 2271  in UTF-8 mode. The code that uses this t Line 2325  in UTF-8 mode. The code that uses this t
2325    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2326    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2327    
2328  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
 condition. */  
2329    
2330  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2331    
# Line 2287  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2340  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2340         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2341         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2342         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2343         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2344           ERR80, ERR81, ERR82, ERR83, ERR84, ERRCOUNT };
2345    
2346  /* JIT compiling modes. The function list is indexed by them. */  /* JIT compiling modes. The function list is indexed by them. */
2347    
2348  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2349         JIT_NUMBER_OF_COMPILE_MODES };         JIT_NUMBER_OF_COMPILE_MODES };
2350    
# Line 2407  typedef struct open_capitem { Line 2462  typedef struct open_capitem {
2462    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2463  } open_capitem;  } open_capitem;
2464    
2465  /* Structure for building a list of named groups during the first pass of  /* Structure for building a list of named groups during the first pass of
2466  compiling. */  compiling. */
2467    
2468  typedef struct named_group {  typedef struct named_group {
2469    const pcre_uchar  *name;          /* Points to the name in the pattern */    const pcre_uchar  *name;          /* Points to the name in the pattern */
2470    int                length;        /* Length of the name */    int                length;        /* Length of the name */
2471    pcre_uint32        number;        /* Group number */    pcre_uint32        number;        /* Group number */
2472  } named_group;  } named_group;
2473    
2474  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2475  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
# Line 2434  typedef struct compile_data { Line 2489  typedef struct compile_data {
2489    pcre_uchar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2490    int  names_found;                 /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2491    int  name_entry_size;             /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2492    int  named_group_list_size;       /* Number of entries in the list */    int  named_group_list_size;       /* Number of entries in the list */
2493    int  workspace_size;              /* Size of workspace */    int  workspace_size;              /* Size of workspace */
2494    unsigned int bracount;            /* Count of capturing parens as we compile */    unsigned int bracount;            /* Count of capturing parens as we compile */
2495    int  final_bracount;              /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
2496    int  max_lookbehind;              /* Maximum lookbehind (characters) */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2497    int  top_backref;                 /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
2498    unsigned int backref_map;         /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
2499      unsigned int namedrefcount;       /* Number of backreferences by name */
2500      int  parens_depth;                /* Depth of nested parentheses */
2501    int  assert_depth;                /* Depth of nested assertions */    int  assert_depth;                /* Depth of nested assertions */
2502    pcre_uint32 external_options;     /* External (initial) options */    pcre_uint32 external_options;     /* External (initial) options */
2503    pcre_uint32 external_flags;       /* External flag bits to be set */    pcre_uint32 external_flags;       /* External flag bits to be set */
# Line 2448  typedef struct compile_data { Line 2505  typedef struct compile_data {
2505    BOOL had_accept;                  /* (*ACCEPT) encountered */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2506    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2507    BOOL check_lookbehind;            /* Lookbehinds need later checking */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
2508      BOOL dupnames;                    /* Duplicate names exist */
2509    int  nltype;                      /* Newline type */    int  nltype;                      /* Newline type */
2510    int  nllen;                       /* Newline string length */    int  nllen;                       /* Newline string length */
2511    pcre_uchar nl[4];                 /* Newline string when fixed length */    pcre_uchar nl[4];                 /* Newline string when fixed length */

Legend:
Removed from v.1359  
changed lines
  Added in v.1409

  ViewVC Help
Powered by ViewVC 1.1.5